We have undertaken a fresh study of this ancient problem, beginning with the necropsy records of Bellevue Hospital in New York and the Massachusetts General Hospital in Boston. In each of these institutions we examined three thousand necropsy records, searching for lesions ordinarily supposed to produce diarrhea. We then traced the cases showing these lesions back to the clinical records, trying to ascertain first whether the lesions actually produced diarrhea, and if so, in what proportion of cases; secondly, whether any special type of symptoms or of discharges was associated with any special lesion of the intestine, the endeavor being to mark out clinical types so far as this was possible. Finally, we reviewed the results of treatment both in necropsy cases and in a considerable series of cases which did not come to necropsy, and endeavored to estimate the value of the different methods used.
For various reasons we have excluded from this study certain